Personally inputs and output are more of a formal thing that's higher in the hierarchy of things to consider, like when I get a project I gather all the information, plans etc at my disposal, those would be the inputs. Then I churn out a 3D model or plans, sections, visuals, video, etc. These are the outputs that my client needs. Sometimes I create outputs that will be used to refine the model, eg spotting an error or an inconsistencies on a plan view that was obscured by some part of the model in the 3D view. But my client doesn't care about these, so it's not really an output of my workflow, rather a part that serves to build the final product.
I like the framing of having a 3D model "ground truth" (albeit taking into account level of detail settings) and being able to derive different views which are designed to highlights some parts of the 3D model or some semantic information on the elements of the model. Automatic annotations are not that hard to do. Meaningful automatic annotations, though... If you're not using a fully constraints based system it's really hard to create an algorithm that will know which properties to highlight. Does my client care that this wall is 30 cm thick, or does he care that its interior surface area is 4m², or does he care that it is made of bricks ?
Comments
Personally inputs and output are more of a formal thing that's higher in the hierarchy of things to consider, like when I get a project I gather all the information, plans etc at my disposal, those would be the inputs. Then I churn out a 3D model or plans, sections, visuals, video, etc. These are the outputs that my client needs. Sometimes I create outputs that will be used to refine the model, eg spotting an error or an inconsistencies on a plan view that was obscured by some part of the model in the 3D view. But my client doesn't care about these, so it's not really an output of my workflow, rather a part that serves to build the final product.
I like the framing of having a 3D model "ground truth" (albeit taking into account level of detail settings) and being able to derive different views which are designed to highlights some parts of the 3D model or some semantic information on the elements of the model. Automatic annotations are not that hard to do. Meaningful automatic annotations, though... If you're not using a fully constraints based system it's really hard to create an algorithm that will know which properties to highlight. Does my client care that this wall is 30 cm thick, or does he care that its interior surface area is 4m², or does he care that it is made of bricks ?